[XeTeX] making a format file of ConTeXt using XeTeX

[Jonathan Kew]

On 7 Mar 2005, at 3:49 pm, Dalyoung wrote:

>> 	ls -l `kpsewhich cmr10.pfb`
>> 	T1Wrap `kpsewhich cmr10.pfb` > /tmp/cmr10.pfb-sfnt.otf
>> 	ls -l /tmp/cmr10.pfb-sfnt.otf
>
> Here are the results of the above commands.
>
> $ ls -l `kpsewhich cmr10.pfb`
> -rw-r--r--  1 mac  unknown  26882  6 Mar  1997 
> /usr/local/teTeX/share/texmf.tetex/fonts/type1/bluesky/cm/cmr10.pfb
>
> $ T1Wrap `kpsewhich cmr10.pfb` > /tmp/cmr10.pfb-sfnt.otf
> 2005-03-08 00:32:13.918 T1Wrap[505] NSScanner: nil string argument
> exception: T1WrapException  reason: expecting only one font definition 
> in pfb (found many or none)

OK, just to make absolutely sure, could you archive and send me a copy 
of your cmr10.pfb file. It should be identical to mine, but let's check 
that it hasn't somehow been corrupted. Please do:
	tar czf ~/cmr10pfb.tgz `kpsewhich cmr10.pfb`
and send me (off-list is fine) the resulting cmr10pfb.tgz archive from 
your home directory.

> From the above outputs, I assume that T1Wrap didn't work properly by 
> some reason which I don't know.
> By the way, why did it look for CM fonts? I never had this kind of 
> errors before the upgrade.

Before version 0.9, XeTeX was using custom .otf versions of the CM 
fonts (they were probably being used by default in some places even if 
you didn't specifically request them). As this was not very robust 
(e.g., the Euler encoding issues), 0.9 attempts to use the standard 
.pfb files by means of an sfnt-format "wrapper". I'm still trying to 
figure out why the T1Wrap tool is failing on your system, when it works 
for others.

JK